HealthTracker fires a torpedo at cardiovascular risk screening
A 40,000-patient strong study into the use of the HealthTracker clinical decision support tool in primary care has shown it can increase screening of patients at high risk of developing cardiovascular disease and contributes new evidence on the effect of technology-assisted interventions to improve healthcare quality.
The Treatment of Cardiovascular Risk using Electronic Decision Support (TORPEDO) study involved 40 metropolitan general practices in Sydney and 20 Aboriginal Community Controlled Health Services (ACCHS) in NSW and Queensland.
Led by Associate Professor David Peiris, head of primary health care research at the George Institute for Global Health, the study looked at how primary care providers could better screen for cardiovascular disease risk and improve prescribing and risk management using the HealthTracker tools.
HealthTracker has been developed by Dr Peiris and his team in association with Pen Computer Systems, and interfaces with MedicalDirector and Best Practice clinical software through Pen's PrimaryCareSidebar tool. It is also fully integrated with Pen's Clinical Audit Tool (CAT) and can provide simple and comprehensive practice performance monitoring and quality improvement.
Published today in the American Heart Association journal Circulation: Cardiovascular Quality and Outcomes, the study's primary outcomes were the proportion of eligible patients who received appropriate screening of CVD risk factors, defined as smoking status, systolic blood pressure, total cholesterol and high-density lipoprotein cholesterol.
It found that the intervention resulted in a 25 per cent relative improvement in screening for cardiovascular disease risk, and importantly from an eHealth perspective, the technology was perceived to be easy to use by clinicians and adds weight to the relatively small amount of published evidence backing the use of clinical decision support tools in primary care.
It also looked at the proportion of eligible patients defined as being at high CVD risk who then received recommended medication prescriptions at the end of study. While this outcome was “a mixed bag”, as Dr Peiris puts it, showing no statistically significant improvement in prescribing patterns for the whole high risk cohort, it did show nearly a 60 per cent relative improvement in prescriptions for recommended medicines and escalated prescribing in the number of people identified at high risk who were not receiving them at the start.
HealthTracker has its origins back in 2007, when the George Institute received a small grant to develop an algorithm that would allow GPs to wade through the growing number of guidelines on chronic disease management at the point of care.
“There were too many conflicting guidelines so we thought what if we could write an algorithm that synthesised all of the key recommendations of the multiple guidelines,” Dr Peiris said. “We basically tested that on some homegrown software and we then got an NHMRC development grant to take us to the next stage.”
The team went out to tender to find a software partner to build a tool that could extract patient data, synthesise the relevant guidelines and produce a graphical representation of risk factors. Pen Computer Systems won that tender, and over the years HealthTracker has been built to be a simple, two- or three-click tool that works with the PrimaryCareSidebar.
Data can also be extracted and audited using the CAT tool to provide a practice-wide snapshot of the particular diagnosis or risk factors being studied. In the TORPEDO trial, the George Institute also worked with the Improvement Foundation (IF) to provide participants with a peer-ranked performance portal for all of the health services participating in the trial.
“There are two components [to HealthTracker],” Dr Peiris said “It has got what we call the in-consultation component, which is what you would use with the patient in front of you. That is currently built into the PrimaryCare Sidebar, and we are migrating it over to Pen's new Top Bar product, which is in beta testing at the moment.
“With the in-consultation component, you select the app from the Sidebar, a couple of mouse clicks, and then there would be a traffic light dashboard of all the inputs and the patient's risk for cardiovascular disease. Then there are recommendations on tests, medication management and meeting guideline targets.”
Another tab in the in-consultation component has a risk communication tool that clinicians can use to discuss risk with the patient, including current and projected risk, and show them the changes in risk if they gave up smoking or they developed diabetes, for example.
“That gives you two printed summaries, either GP version or a patient version, of the report including links to various curated resources, mainly from the Heart Foundation and related peak bodies,” Dr Peiris said. “We tried to design it as something that you could use relatively quickly to alert you to where the gaps might be in the guidelines for that patient and then to discuss that with the patient while they are there.
“The second set of tools we built is what we call the out-of-consultation element and this is the auditing tool. That is the part that is built inside the CAT. With the CAT tool you can do a whole-of-practice snapshot, so we built a new tab inside CAT which only our trial health services got access to.
“That again used graphical displays of practice performance on the key indicators we were interested in. Inside CAT you can bring up the patients that are missing out on a particular thing and then you can set customised prompts in the Sidebar which will pop up every time you open that patient’s record. CAT and the Sidebar were designed to work together.”
For the trial, the team also used a peer-ranked performance portal that functions similarly to that used by the IF as part of its primary care collaboratives program. “We really just put in a few extra bits to what they currently do around looking at where your health service was and how you ranked compared to all the other sites in the trial for six indicators,” he said.
As part of the trial, the George provided the technologies as a whole package and also provided some staff support time.
“One of the key things about this study was we intentionally wanted the support time to be as lean as possible, to have something fairly scalable," Dr Peiris said. "We were quite pleased with that in the sense that it ended up averaging about an hour of support time per month per site.
“Probably about half of that time was at the initial stages in terms of installing the software, password set-up and registering users, but we also made a point of giving everyone a face-to-face training session in how to use the tool and then depending on whether there they wanted to or not, we would then provide either webinar or phone support down the track.”
HealthTracker is also being used in another study Dr Peiris is leading, which is developing a companion consumer portal to HealthTracker. This NHMRC-funded study is using a different data extraction tool – Extensia's RecordConnect, which is compatible with a range of PMSs including MedicalDirector, Best Practice, practiX, statHealth, Genie and Communicare.
“RecordConnect will get the data out into the shared record back-end and we built a front end web app that consumers can log into which will pre-populate with their GP data for diagnoses, medicines, latest test results etc. We've built a whole lot of tools and resources for the patient to hopefully better engage in their care and lower their risks of cardiovascular disease.”
Dr Peiris said the software in this trial seems to be working acceptably for GPs. “It's fairly quiet and sits away in the background and works reasonably unobtrusively. The big question will be whether giving this information to patients leads to better outcomes.”
In addition to measurable health outcomes, much of what Dr Peiris and his team are studying will add to what is relatively limited published evidence in Australia in terms of randomised control trials for using clinical software tools for quality improvement.
Using existing data extraction tools that provide de-identified data for the whole health service was approved by the research ethics committees that reviewed the study.
“The beauty of that is you are not preferentially including the willing and able into a trial – you are getting everyone,” Dr Peiris said. “The Aboriginal Community Controlled Health Services led the way in this study – they were the ones who were most actively interested in participating and are already doing a lot more in quality improvement than mainstream general practice.
“We were really proud of the fact that we had support from the Queensland and NSW state affiliate organisations, and Aboriginal Community Controlled Health Services put their hands up quite readily to be involved.
“What we didn't talk about so much in the paper but is a very hot topic is around the financing of primary care and how we actually pay for quality. This had no financial driver underpinning it – GPs used the tools if they thought they were useful or otherwise didn't – and one of our collaborators, Mark Harris, has done a really nice study showing that this takes time if you want to do good preventative care.
“It adds time to the consultation, maybe as much as 15 minutes of extra time to do it. I think it does help to feed into the current debate around the fee-for-service model and whether we should be looking at blended payments that pay according to quality, and restructuring general practice so that we can see shifts in these sorts of indicators.”
As to the next steps, the plan is to extend the tool to other chronic diseases, including diabetes and chronic kidney disease – which should be available when Pen releases its new Top Bar in the next few months – and then to musculoskeletal and respiratory diseases. Dr Peiris said feedback from GPs indicated they liked HealthTracker and the other tools, but they wanted them to be disease agnostic.
The George also recently signed a $2 million partnership with Telstra Health that might see some new tools developed, he said.
Another key next stage in the program of work is to scale up access to the tools outside of a research setting. Central to this will be to look at the optimum payment models. Dr Peiris said that ideally the users themselves would not be paying for the technology, but rather Primary Health Networks, governments or insurers might pay for the technology on behalf of primary care providers.
“Improved health outcomes for the community at lower cost to the payer is the ideal we are all striving for,” he said.
Posted in Australian eHealth